Loss aversion and cigarette smoking

Cigarette smoking is entirely preventable but contributes to about 480,000 deaths in the United States and 8 million deaths globally each year. The role of personal behaviors in poor health outcomes is now widely acknowledged.

One of the advances leading to this point was the observation of systematic decision-making bias among individuals with substance use disorders (SUDs). For example, individuals with SUD tend to discount delayed rewards more steeply than matched controls. Delay discounting (DD) has since been validated as a reliable predictor of risk for SUD.

The field of behavioral economics has identified several other decision-making biases, but the role these play in SUD has not been studied in any depth.

Loss aversion (LA) refers to the tendency for the experience of a loss to produce a stronger hedonic effect than an equivalent gain.

One?s attitude toward losses might influence their decision to make risky choices for enticing reinforcing outcomes (e.g., the experience of positive psychoactive effects of a drug) despite the risk of a substantial loss (addiction, job loss, other adverse consequences).

While some findings suggest a relationship between LA and SUD, there have been no targeted investigations controlling for confounding factors.

One basic unanswered question for LA, and any potential bias, is the extent to which it is independent of socio-demographic factors (e.g., gender, educational attainment) and other SUD-related decision-making biases, such as DD. The proposed experiment aims to establish the relationship between LA and SUD by addressing three goals.

First is to understand the relationship between LA and vulnerability to cigarette smoking. Second is to study the independence of that association from DD.

Third is to assess the relative strength and independence of LA and DD with other substance use (alcohol, other drugs) as well as other behavioral-health problems (sleep disturbance, depressed mood).

A nationally representative sample of participants will be recruited with the AmeriSpeak panel maintained by NORC at the University of Chicago and complete all study materials online.

Groups of smokers, never-smokers, and former-smokers will be matched on important demographic factors of age, gender, and educational attainment.

LA will be measured with a gamble acceptance task, DD is strongly related to smoking and will be measured to provide a positive control.

The proposed experiment will determine the extent to which LA is related to SUD when controlling for important confounding variables.

Moreover, if LA and DD are independent predictors of SUD risk, this would provide a strong basis to further validate and understand how decision-making biases combine to influence SUD risk. LA might be a protective factor of SUD and this project could identify LA as a novel intervention target.